We propose to continue our prospective investigation of the etiology of non-insulin dependent diabetes (NIDD) with particular attention to diet, obesity, family history, and exercise. The proposed investigation to explore risk factors for diabetes in women takes advantatge of the ongoing prospective cohort study of 121,700 registered nurses ages 40-65 in 1986, who are followed every 2 years by mailed questionnaire. In 1976, 2184 prevalent cases of diabletes were originally reported. Since that time, an increasing number of newly diagnosed cases are being reported. By the end of the 1984 mailing cycle (8 years), approximately 2120 incident cases of diabetes were reported, and thus far, 1013 have been confirmed as NIDD. The proposed study will continue to document and classify the diagnosis of additional new cases arising among this cohort. Information on a number of risk factors for diabetes has been gathered by questionnaire since 1976 including anthropometric data (including waist and hip circumference), exercise, family history and repeated detailed dietary assessment in 1980, 1984 and 1986. In addition, in 1982-83 toenail specimens on approximately 65,000 members of the cohort were collected. Nails from newly diagnosed diabetics ascertained in 1984 and thereafter will be compared with appropriate matched controls for levels of trace minerals which are hypothesized to be associated with etiology of diabetes. Stratified and multivariate techniques will be used to quanitiate the risk of diabetes with different levels of nutrient intake and exericse after controlling for other potentially important risk factors such as age, family history and obesity. Further study of this cohort will yield precise estimates for a number of risk factors, and provide the opportunity to test several dietary hypotheses which arose from our previous observations. A detailed study of body composition among a subsample of the population will permit an accurate quantification of the risk associated with obesity.